<s>
GOFAI	B-General_Concept
is	O
an	O
acronym	O
for	O
"	O
Good	B-General_Concept
Old-Fashioned	I-General_Concept
Artificial	I-General_Concept
Intelligence	I-General_Concept
"	O
invented	O
by	O
the	O
philosopher	O
John	O
Haugeland	O
in	O
his	O
1985	O
book	O
,	O
Artificial	O
Intelligence	O
:	O
The	O
Very	O
Idea	O
.	O
</s>
<s>
Technically	O
,	O
GOFAI	B-General_Concept
refers	O
only	O
to	O
a	O
restricted	O
kind	O
of	O
symbolic	B-General_Concept
AI	I-General_Concept
,	O
namely	O
rule-based	O
or	O
logical	O
agents	O
.	O
</s>
<s>
This	O
approach	O
was	O
popular	O
in	O
the	O
1980s	O
,	O
especially	O
as	O
an	O
approach	O
to	O
implementing	O
expert	B-Application
systems	I-Application
,	O
but	O
symbolic	B-General_Concept
AI	I-General_Concept
has	O
since	O
been	O
extended	O
in	O
many	O
ways	O
to	O
better	O
handle	O
uncertain	O
reasoning	O
and	O
more	O
open-ended	O
systems	O
.	O
</s>
<s>
Some	O
of	O
these	O
extensions	O
include	O
probabilistic	O
reasoning	O
,	O
non-monotonic	O
reasoning	O
,	O
multi-agent	O
systems	O
,	O
and	O
neuro-symbolic	B-General_Concept
systems	I-General_Concept
.	O
</s>
<s>
Significant	O
contributions	O
of	O
symbolic	B-General_Concept
AI	I-General_Concept
,	O
not	O
encompassed	O
by	O
the	O
GOFAI	B-General_Concept
view	O
,	O
include	O
search	B-Application
algorithms	I-Application
;	O
automated	B-Application
planning	I-Application
and	I-Application
scheduling	I-Application
;	O
constraint-based	B-Application
reasoning	I-Application
;	O
the	O
semantic	O
web	O
;	O
ontologies	O
;	O
knowledge	B-Application
graphs	I-Application
;	O
non-monotonic	O
logic	O
;	O
circumscription	B-Application
;	O
automated	B-Application
theorem	I-Application
proving	I-Application
;	O
and	O
symbolic	B-Algorithm
mathematics	I-Algorithm
.	O
</s>
<s>
For	O
a	O
more	O
complete	O
list	O
,	O
see	O
the	O
main	O
article	O
on	O
symbolic	B-General_Concept
AI	I-General_Concept
.	O
</s>
<s>
Although	O
the	O
term	O
GOFAI	B-General_Concept
encompasses	O
only	O
a	O
small	O
part	O
of	O
symbolic	B-General_Concept
AI	I-General_Concept
,	O
prominent	O
in	O
the	O
1980s	O
,	O
contemporary	O
critics	O
of	O
symbolic	B-General_Concept
AI	I-General_Concept
sometimes	O
use	O
GOFAI	B-General_Concept
as	O
a	O
synonym	O
for	O
it	O
.	O
</s>
<s>
This	O
conflation	O
of	O
terms	O
can	O
lead	O
to	O
conclusions	O
that	O
symbolic	B-General_Concept
AI	I-General_Concept
research	O
ended	O
in	O
the	O
1980s	O
and	O
avoided	O
machine	O
learning	O
.	O
</s>
<s>
During	O
the	O
Second	O
AI	O
Summer	O
,	O
i.e.	O
,	O
the	O
expert	B-Application
systems	I-Application
boom	O
of	O
the	O
1980s	O
,	O
production-rule	O
systems	O
requiring	O
knowledge	O
engineering	O
were	O
used	O
to	O
implement	O
expert	B-Application
systems	I-Application
.	O
</s>
<s>
GOFAI	B-General_Concept
correctly	O
describes	O
this	O
approach	O
.	O
</s>
<s>
The	O
Second	O
AI	O
Winter	O
occurred	O
after	O
the	O
expert	B-Application
systems	I-Application
and	O
Lisp	B-Operating_System
Machines	I-Operating_System
markets	O
collapsed	O
.	O
</s>
<s>
Expert	B-Application
systems	I-Application
did	O
not	O
handle	O
uncertainty	O
well	O
,	O
required	O
more	O
resources	O
to	O
build	O
and	O
maintain	O
than	O
expected	O
,	O
and	O
proved	O
brittle	O
outside	O
their	O
intended	O
domains	O
due	O
to	O
a	O
lack	O
of	O
common-sense	O
reasoning	O
capabilities	O
.	O
</s>
<s>
Language-specific	O
Lisp	B-Operating_System
machines	I-Operating_System
could	O
become	O
easily	O
replaced	O
by	O
newer	O
workstations	O
with	O
similar	O
performance	O
,	O
obviating	O
their	O
need	O
.	O
</s>
<s>
Symbolic	B-General_Concept
AI	I-General_Concept
continued	O
,	O
albeit	O
with	O
reduced	O
funding	O
.	O
</s>
<s>
Work	O
in	O
semantic	O
networks	O
and	O
knowledge	O
representations	O
led	O
to	O
formalizing	O
ontologies	O
with	O
languages	O
such	O
as	O
RDF	O
and	O
OWL	B-Language
,	O
leading	O
to	O
large	O
ontologies	O
such	O
as	O
YAGO	O
.	O
</s>
<s>
However	O
,	O
the	O
research	O
and	O
applications	O
of	O
symbolic	B-General_Concept
AI	I-General_Concept
since	O
the	O
Second	O
AI	O
Winter	O
and	O
outside	O
of	O
the	O
production-rule	O
approach	O
to	O
expert	B-Application
systems	I-Application
are	O
less	O
well	O
known	O
and	O
now	O
eclipsed	O
by	O
the	O
media	O
focus	O
on	O
deep	B-Algorithm
learning	I-Algorithm
since	O
2012	O
.	O
</s>
<s>
For	O
example	O
,	O
research	O
in	O
symbolic	B-General_Concept
AI	I-General_Concept
machine	O
learning	O
includes	O
decision	B-Algorithm
tree	I-Algorithm
learning	I-Algorithm
,	O
Mitchell	O
's	O
version	B-General_Concept
space	I-General_Concept
learning	I-General_Concept
,	O
Valiant	O
's	O
contributions	O
to	O
PAC	O
learning	O
,	O
statistical	B-Algorithm
relational	I-Algorithm
learning	I-Algorithm
,	O
inductive	B-Application
logic	I-Application
programming	I-Application
,	O
and	O
various	O
interactive	O
learning	O
approaches	O
to	O
incorporate	O
user	O
advice	B-Application
,	O
examples	O
,	O
and	O
explanations	O
as	O
an	O
integral	O
part	O
of	O
the	O
learning	O
process	O
.	O
</s>
<s>
Using	O
GOFAI	B-General_Concept
as	O
a	O
synonym	O
for	O
current	O
symbolic	B-General_Concept
AI	I-General_Concept
leads	O
to	O
erroneous	O
conclusions	O
and	O
confusion	O
.	O
</s>
<s>
Garcez	O
and	O
Lamb	O
provide	O
an	O
example:Turing	O
award	O
winner	O
Judea	O
Pearl	O
offers	O
a	O
critique	O
of	O
machine	O
learning	O
which	O
,	O
unfortunately	O
,	O
conflates	O
the	O
terms	O
machine	O
learning	O
and	O
deep	B-Algorithm
learning	I-Algorithm
.	O
</s>
<s>
Similarly	O
,	O
when	O
Geoffrey	O
Hinton	O
refers	O
to	O
symbolic	B-General_Concept
AI	I-General_Concept
,	O
the	O
connotation	O
of	O
the	O
term	O
tends	O
to	O
be	O
that	O
of	O
expert	B-Application
systems	I-Application
dispossessed	O
of	O
any	O
ability	O
to	O
learn	O
.	O
</s>
<s>
Machine	O
learning	O
is	O
not	O
confined	O
to	O
association	B-Algorithm
rule	I-Algorithm
mining	I-Algorithm
,	O
c.f.	O
</s>
<s>
the	O
body	O
of	O
work	O
on	O
symbolic	O
ML	O
[	O
machine	O
learning ]	O
and	O
relational	B-Application
learning	I-Application
(	O
the	O
differences	O
to	O
deep	B-Algorithm
learning	I-Algorithm
being	O
the	O
choice	O
of	O
representation	O
,	O
localist	O
logical	O
rather	O
than	O
distributed	O
,	O
and	O
the	O
non-use	O
of	O
gradient-based	O
learning	O
algorithms	O
)	O
.	O
</s>
<s>
Equally	O
,	O
symbolic	B-General_Concept
AI	I-General_Concept
is	O
not	O
just	O
about	O
production	B-Application
rules	I-Application
written	O
by	O
hand	O
.	O
</s>
<s>
A	O
proper	O
definition	O
of	O
AI	O
concerns	O
knowledge	O
representation	O
and	O
reasoning	O
,	O
autonomous	O
multi-agent	O
systems	O
,	O
planning	B-Application
and	O
argumentation	B-General_Concept
,	O
as	O
well	O
as	O
learning	O
.	O
</s>
<s>
The	O
key	O
points	O
above	O
are	O
that	O
symbolic	B-General_Concept
AI	I-General_Concept
research	O
has	O
long	O
since	O
moved	O
beyond	O
GOFAI	B-General_Concept
,	O
research	O
continues	O
,	O
and	O
GOFAI	B-General_Concept
no	O
longer	O
describes	O
it	O
.	O
</s>
<s>
Further	O
,	O
there	O
are	O
symbolic	O
learning	O
approaches	O
to	O
machine	O
learning	O
,	O
such	O
as	O
inductive	B-Application
logic	I-Application
programming	I-Application
and	O
statistical	B-Algorithm
relational	I-Algorithm
learning	I-Algorithm
,	O
i.e.	O
,	O
it	O
is	O
not	O
just	O
the	O
domain	O
of	O
deep	B-Algorithm
learning	I-Algorithm
.	O
</s>
<s>
Below	O
,	O
we	O
return	O
to	O
the	O
philosophical	O
critiques	O
leveled	O
against	O
GOFAI	B-General_Concept
,	O
the	O
symbolic	B-General_Concept
AI	I-General_Concept
approach	O
of	O
the	O
1980s	O
.	O
</s>
<s>
GOFAI	B-General_Concept
,	O
the	O
rule-based	O
approach	O
of	O
1980s	O
symbolic	B-General_Concept
AI	I-General_Concept
,	O
was	O
attacked	O
by	O
philosophers	O
such	O
as	O
Hubert	O
Dreyfus	O
,	O
his	O
brother	O
Stuart	O
Dreyfus	O
,	O
and	O
philosopher	O
Kenneth	O
Sayre	O
.	O
</s>
<s>
Russell	O
and	O
Norvig	O
,	O
describe	O
the	O
GOFAI	B-General_Concept
critique	O
in	O
Artificial	O
Intelligence	O
:	O
A	O
Modern	O
Approach:The	O
position	O
they	O
criticize	O
came	O
to	O
be	O
called	O
"	O
Good	O
Old-Fashioned	O
Al	O
,	O
"	O
or	O
GOFAI	B-General_Concept
,	O
a	O
term	O
coined	O
by	O
Haugeland	O
(	O
1985	O
)	O
.	O
</s>
<s>
GOFAI	B-General_Concept
is	O
supposed	O
to	O
claim	O
that	O
all	O
intelligent	O
behavior	O
can	O
be	O
captured	O
by	O
a	O
system	O
that	O
reasons	O
logically	O
from	O
a	O
set	O
of	O
facts	O
and	O
rules	O
describing	O
the	O
domain	O
.	O
</s>
<s>
It	O
is	O
reasonable	O
to	O
suppose	O
,	O
however	O
,	O
that	O
a	O
book	O
called	O
What	O
First-Order	O
Logical	O
Rule-Based	O
Systems	O
Without	O
Learning	O
Ca	O
n't	O
Do	O
might	O
have	O
had	O
less	O
impact.In	O
other	O
words	O
,	O
GOFAI	B-General_Concept
restricts	O
its	O
view	O
of	O
agents	O
to	O
those	O
controlled	O
by	O
logical	O
rules	O
.	O
</s>
<s>
In	O
contrast	O
to	O
this	O
view	O
,	O
symbolic	B-General_Concept
AI	I-General_Concept
also	O
includes	O
non-monotonic	O
logic	O
,	O
modal	O
logic	O
,	O
probabilistic	O
logics	O
,	O
multi-agent	O
systems	O
,	O
symbolic	O
machine	O
learning	O
,	O
and	O
hybrid	O
neuro-symbolic	B-General_Concept
architectures	I-General_Concept
.	O
</s>
<s>
Symbolic	O
machine	O
learning	O
,	O
i.e.	O
,	O
non-connectionist	O
machine	O
learning	O
specific	O
to	O
symbolic	B-General_Concept
AI	I-General_Concept
,	O
includes	O
inductive	B-Application
logic	I-Application
programming	I-Application
,	O
statistical	B-Algorithm
relational	I-Algorithm
learning	I-Algorithm
,	O
case-based	O
learning	O
,	O
knowledge	O
compilation	O
(	O
chunking	O
)	O
,	O
macro-operator	O
learning	O
,	O
learning	O
from	O
analogy	O
,	O
and	O
interactive	O
learning	O
from	O
human	O
advice	B-Application
,	O
explanations	O
,	O
and	O
exemplars	O
.	O
</s>
<s>
Russell	O
and	O
Norvig	O
do	O
not	O
reject	O
all	O
of	O
Dreyfus	O
’s	O
arguments	O
,	O
they	O
accept	O
his	O
strongest	O
argument	O
,	O
one	O
that	O
applies	O
to	O
all	O
disembodied	O
AIs	O
,	O
whatever	O
their	O
approach:One	O
of	O
Dreyfus	O
's	O
strongest	O
arguments	O
is	O
for	O
situated	B-General_Concept
agents	I-General_Concept
rather	O
than	O
disembodied	O
logical	O
inference	O
engines	O
.	O
</s>
